Critical Thinking Education and Debiasing TIM KENYON University of Waterloo Department of Philosophy 200 University Avenue West Waterloo, ON N2L 3G1 Canada [email protected] GUILLAUME BEAULAC Yale University Department of Philosophy 344 College St. New Haven, CT. 06511-6629 USA [email protected] Abstract: There are empirical Résumé: Des données empiriques grounds to doubt the effectiveness of nous permettent de douter de a common and intuitive approach to l'efficacité d'une approche commune teaching debiasing strategies in et intuitive pour enseigner des critical thinking courses. We stratégies de correction de biais summarize some of the grounds cognitifs dans les cours de pensée before suggesting a broader critique. Nous résumons certains de taxonomy of debiasing strategies. ces résultats empiriques avant de This four-level taxonomy enables a suggérer une taxonomie plus useful diagnosis of biasing factors étendue de ces stratégies de and situations, and illuminates more correction de biais. Cette taxonomie strategies for more effective bias à quatre niveaux permet un mitigation located in the shaping of diagnostic utile de facteurs causant situational factors and reasoning les biais et elle met en évidence infrastructure—sometimes called davantage de stratégies permettant la “nudges” in the literature. The correction plus efficace de biais, question, we contend, then becomes stratégies situées dans des mesures how best to teach the construction modifiant les infrastructures et les and use of such infrastructures. environnements cognitifs ("nudge" dans la littérature). Nous soutenons que la question porte dès lors sur les meilleures façons d'enseigner la construction et l'utilisation de ces infrastructures. Keywords: Critical thinking, biases, debiasing, education, nudges © Tim Kenyon and Guillaume Beaulac, Informal Logic, Vol. 34, No. 4 (2014), pp. 341-363. 342 Tim Kenyon and Guillaume Beaulac 1. Introduction Teaching critical thinking is an undertaking that permits emphasis on many different combinations of elements, the most traditional of which are formal logic, informal logic, argumentation, fallacy theory, and rhetoric. Increasingly, too, critical thinking courses and texts include an explicit emphasis on the psychology of cognitive and social biases (see, for example, Kenyon 2008; Ruggiero 2004; Groarke & Tindale 2004; and Gilovich 1991). While they vary greatly in the length and detail of their treatments, a common feature of these texts is that they present names, taxonomies and definitions of some key biases, perhaps with some examples or explanations of the underlying empirical work included. Given their role in critical thinking didactics, it is safe to assume that these treatments are intended to foster practical reasoning skills of mitigating or forestalling the effects of biases – to enable students to identify biases in reasoning, and to minimize biases in their own thinking. The overall aim is consonant with the general rationale for teaching critical reasoning courses in the first place. Yet these texts also commonly lack empirically-informed material, distinct from that already mentioned, that aims to teach students the skills of minimizing bias in their thinking or their actions. In other words, the combination of what such treatments do and do not contain reflects the assumption that simply teaching students about biases is an effective way of enabling them to reduce the distortions of biases in their own thinking. We identify this assumption as the intuitive approach to teaching debiasing, or IA. (IA) Teaching facts about biases, including a taxonomy of biases and their various propensities to distort reasoning, is a reasonably effective means of providing students in critical reasoning courses with skills enabling the detection and mitigation of biases, including students’ own biases. Something along the lines of IA informs the treatments that biases receive in the critical thinking texts already noted. It is also central to some reviews on the topic (e.g., Larrick 2004), while its influence can be seen also in training contexts beyond that of a critical thinking course. An example of this latter type of context is Croskerry, Singhal, & Mamede’s (2013a, 2013b) © Tim Kenyon and Guillaume Beaulac, Informal Logic, Vol. 34, No. 4 (2014), pp. 341-363. Critical Thinking Education and Debiasing 343 approach to cognitive debiasing for clinicians’ medical judgments: even though Croskerry et al. record a “general pessimism […] about the feasibility of cognitive debiasing” (Croskerry et al. 2013a, p. ii63), they adopt the recommendation that clinicians “must be informed and recognise the need for constant vigilance and surveillance of their [own] thinking to mitigate diagnostic and other clinical errors” (Croskerry et al. 2013b, p. 6).1 It is noteworthy that IA characterizes much of how critical thinking education treats debiasing, we contend, because when one considers the empirical evidence bearing on it, the most plausible simple view of IA is that it is—at least in most cases— false.2 At least, the practice of simply teaching students facts about biases is not as effective as one might hope. The literature on the cognitive and social psychology of debiasing indicates, on balance, that teaching people about biases does not reliably debias them. Indeed, the literature suggests that (for at least a wide class of biases) practically any debiasing strategy intended to be learned and subsequently self-deployed by individuals, acting alone and at the point of making a judgment, is unlikely to succeed in significantly minimizing biases. In the following remarks, we briefly outline why this is so before moving on to consider the ramifications for critical thinking education. Vast resources are currently devoted to teaching critical reasoning worldwide. Does the implausibility of IA mean that these resources are misused, to the extent that they are predicated on IA? Should philosophers, psychologists, and other critical reasoning educators just stop including a focus on biases in critical thinking education? We do not think so. Rather, we take the lesson to be that whole societies and polities have a major interest in promoting efficacious debiasing education—extending to population-level demographic scales and intergenerational time scales. The difficulty of teaching debiasing skills that could be deployed in a strictly atomistic or individualistic way counts in favor of teaching and investing also in more collective debiasing strategies and infrastructure that would serve the latter sorts of interests. This approach will encompass teaching not just individual skills and knowledge, but skills that enable the 1 In fact the distinctions between approaches to critical thinking that we propose in the following remarks should be helpful in characterizing the kinds of clinical training strategies described by Croskerry et al. (2013a, 2013b). 2 Below, we identify some methods that would fall under IA that we believe to be relatively promising. © Tim Kenyon and Guillaume Beaulac, Informal Logic, Vol. 34, No. 4 (2014), pp. 241-263. 344 Tim Kenyon and Guillaume Beaulac construction of reasoning infrastructure, and effective participation in social and organizational reasoning processes and decision procedures. What would these processes, strategies and infrastructure look like? A key first step here is to reflect on the breadth of what can count as debiasing from a critical thinking perspective. Our aim in this reflection is to help motivate and set the stage for creative and empirically guided work on how to teach debiasing in ways that might be efficacious, serving both private and public interests in minimizing distorted or unreliable reasoning. By focusing on choices, behavior, and agent-world interactions, we suggest a broader range of outcomes for critical thinking than that informing IA, and therefore a broader range of options for critical thinking education as well. Those familiar with the critical thinking literature can skip section 2, in which we develop and justify our characterization of IA. In section 3, we present some reasons for pessimism towards IA. Finally, in sections 4 and 5, we introduce our positive proposal by first distinguishing different ways in which we can debias and then discussing how this impacts the way we conceive critical thinking education. 2. Characterizing the intuitive approach First, an explanation and a caveat. By ‘bias’ we most generally mean something neutral with respect to both moral properties and questions of accuracy. A bias in this sense is simply a disposition, implicit or explicit, to reach a particular kind of conclusion or outcome, or to remain in one. This interpretation, common in the psychological literature, accommodates the idea that biases can skew a process in a way that makes its outcome inaccurate or otherwise wrong, but it also leaves open the prospect that biases play a role in truth-conducive reasoning processes and morally unproblematic judgments or attitudes. In common parlance, of course, one normally goes to the trouble of saying that some attitude, reasoning, or person is biased only if the operation of the bias is claimed to be problematic—a distortion, or a prejudice that amounts to a vice. Our focus on debiasing is one that presumes the former meaning: it is predicated on the thought that biases should be mitigated when they are problematic, and not because they are by definition problematic (e.g., Klein & Kahneman 2009 explore when and